Smart Contracts for Spectrum Sensing as a Service

Abstract

Mobile network operators (MNO) can opportunistically use the licensed bands of the primary users (PU) provided that they monitor the spectrum and stop their transmission upon detection of the PU. As deploying spectrum sensors may be prohibitively expensive, the MNO can buy spectrum sensing service from sensing helpers in its proximity. However, such a trade requires a framework with three key functions: helper selection, faulty or malicious helper identification, and payment to honest helpers. Here, we introduce Spass which provides these functions and facilitates a fair exchange between the entities without a trusted third party via smart contracts (SC) running on a blockchain network. While payments via SCs seem conceptually simple, realizing it is difficult due to the cost of using SC functions which might be prohibitive as write/computation operations on the SCs might have a cost, e.g., in Ethereum. Considering our design goals and SC-related overhead, we derive the optimal Spass parameters maximizing the MNO's profit. Moreover, we propose a K-means clustering approach to identify independent malicious helpers, and using both lossless and lossy compression on the helpers' sensing report to decrease the cost of write operations. Via simulations, we show under which conditions Spass-powered service leads to a profitable business for an MNO.